/metric-learning

Primary LanguageMATLABMIT LicenseMIT

metric-learning

Faster implementation of algorithms for metric-learning. As Wiki describes metric-learning as task of learning distance function over objects. They follow following properties:

  • Non-negativity - greater than or equal to 0
  • Symmetry
  • Triangle inequality

some of the algorithms for metric-learning include

  1. large-margin nearest neighbor
  2. information-theoretic metric learning

##References

[1]Jason V. Davis, Brian Kulis, Prateek Jain, Suvrit Sra, and Inderjit S. Dhillon. "Information-theoretic Metric Learning." Proc. 24th International Conference on Machine Learning (ICML), 2007.